Laurence Devillers and the Gender Bias in Artificial Intelligence

By Sophie Videment

There has been an intense debate lately on the hopes and fears raised by the use of Artificial Intelligence in all aspects of our daily lives, including our jobs, homes, and vehicles.

While research on Artificial Intelligence dates back to the 1960s, there have been spectacular improvements in recent years. The applications of AI have multiplied and a real technological revolution is announced: autonomous vehicles, medical diagnostics, algorithmic finance, industrial robotics and intelligent weapons,** to name but a few.

Having researched numerous articles and videos on the subject, I am particularly concerned by an article authored by the French researcher Laurence Devillers * with one point being especially relevant to our concern at PWN: the gender bias which may be perpetuated by algorithms and machine learning. M. Devillers concludes that, with AI and Big Data, behavior models are programmed to predict future behaviors. So, when they are included in chatbots on the internet or embarked in robots to sense, detect and decide, these models often contain bias, related to the decision choices process made by the engineers who have designed them. If robots learn by themselves on the basis of data collected on their own, without intervention, these biasses will reflect those existing of our society.

Today, as we know, women appear less often than men in movies, media, conferences and companies Executive boards. Also, a large majority of the professionals of the Tech giants are male. A lack of diversity in the AI and digital industry naturally reinforces these gender stereotypes. If the humans who teach computers how to behave are male only, robots will very likely reflect the programmer’s vision of the world.

What can be done to rectify this?  Be mindful, forceful and give voice to the question of gender bias within the society debates on the ethics and AI.  This is also a very strong reason to entice and promote the number of women to enter the Tech industry.

*Laurence Devillers is the author of « Des robots et des hommes : mythes, fantasmes et réalité », Plon 2017

**If you have missed it: Physicist Stephen Hawking and Tesla CEO Elon Musk along with hundreds of other experts have publicly shown their concern this summer about the dangers of the military use of AI in an open letter addressed to the United Nations, which lead to an intense controversy between Elon Musk and Mark Zuckerberg.